arXiv subjects

Fenwick C. Cooper

Publications and source records attributed to Fenwick C. Cooper.

4 recordsLinked to original sources

arXiv2025

Improved rainfall forecasts in daily use over East Africa

Ensemble forecasting has proven to be a vital tool for predicting extreme, life-threatening or only partially predictable weather events. As well as providing probabilistic products, individual ensemble members provide context and realisations of possible extreme weather. However, many National Meteorological Services in East Africa do not have the computing resources to enable them to run their local area models in ensemble mode over the full period of the two-week medium range. In this paper we test the performance of a forecast system, comprising the global ECMWF ensemble forecast, post-processed using cGAN, a neural network model, and forecasts calibrated using IDR, a recent statistical method, against the probabilistic climatology. cGAN provides comparable levels of probabilistic skill as IDR applied to the ECMWF ensemble or IDR applied to the machine-learning models FuXi and GraphCast. These methods improve the raw ECMWF ensemble forecast substantially, which is itself an improvement on deterministic forecasts. The ability of cGAN to produce individual realisations of future rainfall was important when assessed in an operational context by the National Meteorological and Hydrological Services of Kenya and Ethiopia. Moreover, in common with IDR, cGAN is cheap to train/run and requires no additional post-processing. It is run on laptops to generate many thousands of ensemble members, making it suitable for Meteorological Services with limited computational facilities.

physics.ao-ph↗

arXiv2023

Further analysis of cGAN: A system for Generative Deep Learning Post-processing of Precipitation

The conditional generative adversarial rainfall model "cGAN" developed for the UK \cite{Harris22} was trained to post-process into an ensemble and downscale ERA5 rainfall to 1km resolution over three regions of the USA and the UK. Relative to radar data (stage IV and NIMROD), the quality of the forecast rainfall distribution was quantified locally at each grid point and between grid points using the spatial correlation structure. Despite only having information from a single lower quality analysis, the ensembles of post processed rainfall produced were found to be competitive with IFS ensemble forecasts with lead times of between 8 and 16 hours. Comparison to the original cGAN trained on the UK using the IFS HRES forecast indicates that improved training forecasts result in improved post-processing. The cGAN models were additionally applied to the regions that they were not trained on. Each model performed well in their own region indicating that each model is somewhat region specific. However the model trained on the Washington DC, Atlantic coast, region achieved good scores across the USA and was competitive over the UK. There are more overall rainfall events spread over the whole region so the improved scores might be simply due to increased data. A model was therefore trained using data from all four regions which then outperformed the models trained locally.

physics.ao-ph↗

arXiv2015

High dimensional linear inverse modelling

We introduce and demonstrate two linear inverse modelling methods for systems of stochastic ODE's with accuracy that is independent of the dimensionality (number of elements) of the state vector representing the system in question. Truncation of the state space is not required. Instead we rely on the principle that perturbations decay with distance or the fact that for many systems, the state of each data point is only determined at an instant by itself and its neighbours. We further show that all necessary calculations, as well as numerical integration of the resulting linear stochastic system, require computational time and memory proportional to the dimensionality of the state vector.

physics.data-an↗

arXiv2014

Optimisation of an idealised ocean model, stochastic parameterisation of sub-grid eddies

An optimisation scheme is developed to accurately represent the sub-grid scale forcing of a high dimensional chaotic ocean system. Using a simple parameterisation scheme, the velocity components of a 30km resolution shallow water ocean model are optimised to have the same climatological mean and variance as that of a less viscous 7.5km resolution model. The 5 day lag-covariance is also optimised, leading to a more accurate estimate of the high resolution response to forcing using the low resolution model. The system considered is an idealised barotropic double gyre that is chaotic at both resolutions. Using the optimisation scheme, we find and apply the constant in time, but spatially varying, forcing term that is equal to the time integrated forcing of the sub-mesoscale eddies. A linear stochastic term, independent of the large-scale flow, with no spatial correlation but a spatially varying amplitude and time scale is used to represent the transient eddies. The climatological mean, variance and 5 day lag-covariance of the velocity from a single high resolution integration is used to provide an optimisation target. No other high resolution statistics are required. Additional programming effort, for example to build a tangent linear or adjoint model, is not required either. The focus of this paper is on the optimisation scheme and the accuracy of the optimised flow. The method can be applied in future investigations into the physical processes that govern barotropic turbulence and it can perhaps be applied to help understand and correct biases in the mean and variance of a more realistic coarse or eddy-permitting ocean model. The method is complementary to current parameterisations and can be applied at the same time without modification.

physics.ao-ph↗